YAML Metadata Warning:empty or missing yaml metadata in repo card
Check out the documentation for more information.
DEPRECATED — moved to avgrocketman/compile-cache-t4
2026-11-06: This repo is a read-only redirect. New runs use the chip-scoped
cache repo avgrocketman/compile-cache-t4.
| Old path | New path |
|---|---|
llama-dory-t4-cuda.tar.gz |
compile-cache-t4/llama.cpp/98a74c97ea09ac72f6ceda1559808fc0a2f447ce/sm75/cuda13.0/llama-server-t4-sm75-cuda13.0.tar.gz |
The archive is byte-identical (SHA-256
d58c4433aabe5a500b800c1949498591e94f3e887c5f21e6729308633524e5d5), so existing
pins still verify. No new artifacts will be uploaded here.
The new repo contains the shared T4 llama.cpp server for Dory1.0-4B and
Spark-X2.6-4B, the full build manifest, and per-model manifests. It is still
sm_75-only; do not use it on L4/A100/G4.
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support